Market Analysis • February 12, 2026
Claims Down, Trend Up: Feb 12 Release Hides a 7,000 Jump in the 4-Week Average
The Department of Labor’s weekly claims report dated February 12, 2026 leads with a neat headline: seasonally adjusted initial claims fell 5,000 to 227,000. But the trend is doing the opposite. The 4-week average rose by 7,000 to 219,500, revisions nudged last week’s claims up by 1,000 (to 232,000), and continuing claims climbed 21,000 to 1,862,000 even as the insured unemployment rate held at 1.2%. The “stable rate” story sits awkwardly beside a rising stock of unemployed.
Here’s what the data reveals:
- Initial claims fell to 227,000, but the 4-week average jumped to 219,500 and prior week claims were revised up to 232,000
- Continuing claims rose 21,000 to 1,862,000, while the insured unemployment rate stayed at 1.2% for the 10th straight week
- The 4-week average of insured unemployment dipped to 1,846,750 (lowest since October 5, 2024)—a smoothing effect that masks the latest weekly increase
- Seasonal models misfired: NSA initial claims fell by 4,555 when an increase of 1,161 was expected; NSA insured unemployment rose by 3,455 when a decline of 22,089 was expected
- All programs show stress: Total continued weeks claimed rose 76,822 to 2,248,314 (week ending January 24); Short-Time Compensation increased 2,326 to 23,659
- Year-over-year, NSA initial claims and insured unemployment are both higher (248,397 vs 232,745 and 2,214,205 vs 2,187,841) while the NSA insured unemployment rate holds at 1.4%
The Headline Decline Meets a Rising Baseline
The single-week dip to 227,000 reads like relief. The trendline disagrees. The 4-week average at 219,500 is up 7,000 week over week, and the prior week was revised up to 232,000. That’s not noise; it’s the definition of a turn.
Monthly sequencing underscores the drift:
- Initial claims (SA): 199k (Jan 10), 210k (Jan 17), 209k (Jan 24), then 232k (Jan 31) and 227k (Feb 7)
- The 4-week average rose from 204.0k (Jan 17) to 212.5k (Jan 31) to 219.5k (Feb 7)
Call it what it is: post-holiday volatility is fading, and claims have edged higher off early-January troughs.
| Metric | Latest | WoW Change | 4-Week Avg | Prior-Week Revision |
|---|---|---|---|---|
| Initial claims (SA) | 227,000 | -5,000 | 219,500 | +1,000 (231k→232k) |
| Insured unemployment (SA) | 1,862,000 | +21,000 | 1,846,750 | -3,000 (1,844k→1,841k) |
| Insured unemployment rate (SA) | 1.2% | 0.0 pp | — | — |
| All programs continued weeks | 2,248,314 | +76,822 | — | — |
| STC/Workshare | 23,659 | +2,326 | — | — |
The table explains the contradiction: the level data is drifting up while the chosen headline metric flatters the picture.
Continuing Claims: Rate Frozen, Stock Rising
An unchanged 1.2% insured unemployment rate for the week ending January 31 keeps the optics calm, but the count moved the wrong way: +21,000 to 1,862,000. The 4-week average of insured unemployment fell to 1,846,750, the lowest since October 5, 2024—because smoothing removed earlier higher weeks, not because the latest week improved.
The rate has held at 1.2% for 10 straight weeks (Nov 29, 2025 through Jan 31, 2026). Stability in the ratio is masking an upward drift in the numerator. That’s the kind of “quiet turn” markets routinely miss until it shows up in payrolls or consumer delinquencies.
Seasonal Factors vs. Actual Flows
Seasonal factors expected a small rise in unadjusted initial claims; reality delivered a 4,555 drop instead. Good news, but not decisive. Why? Because the unadjusted insured unemployment rose by 3,455 when seasonal factors looked for a 22,089 drop. Under the hood, fewer people filed first-time claims than expected—but more people are staying on the rolls than models implied. That is classic late-cycle friction.
The math reinforces it:
- NSA initial claims: 248,397, up versus the comparable week in 2025 (232,745)
- NSA insured unemployment: 2,214,205 vs 2,187,841 a year ago
- NSA insured unemployment rate: unchanged at 1.4%, but with a higher base
In short, momentum is weaker than the headline suggests.
Geography Matters: Stress Pockets Behind the Aggregate
The national rate blurs persistent regional heat:
- Highest insured unemployment rates: Rhode Island (2.9), New Jersey (2.8), Massachusetts (2.7), Minnesota (2.6), Washington (2.6), with California (2.3), Illinois (2.2), Montana (2.2), New York (2.1), Connecticut (2.0), Michigan (2.0), Oregon (2.0) trailing closely
Weekly flows confirm concentration and volatility:
- Largest increases in initial claims (week ending Jan 31): Pennsylvania (+5,268), New York (+3,141), Missouri (+2,833), New Jersey (+2,602), Illinois (+2,203)
- Insured unemployment increases (advance, week ending Jan 31): New York (+14,291), Pennsylvania (+10,765), New Jersey (+6,347), Washington (+5,875), Oregon (+3,250)
- Offsets from earlier declines: California (-23,061), Texas (-11,916), Florida (-5,702)
- The latest advance initial claims (week ending Feb 7) then flipped some of these: Texas (+2,185), Virginia (+1,845), Pennsylvania (-3,287), Missouri (-2,948), Illinois (-2,349)
Translation: state-level tightness and loosening are ping-ponging week to week, but stress remains clustered in the Northeast and West Coast. For credit and muni investors, that geographic skew matters.
Monthly Drift and the Revision Fog
We’ve seen this movie: the release leans into a “lowest 4-week insured average since 2024” angle while the latest week’s continuing claims climb. Revisions then reshape the path: initial claims revised up +1,000, continuing claims revised down -3,000. This persistent pattern influences the moving averages and subtly reframes inflection points. The broader arc since mid-January is clear enough: cooling labor resilience with a gentle firming in both initial and continuing claims off the troughs.
What This Means for Markets
- Rates: A deceptively soft headline with a firming trend is classic “don’t overreact” territory. The drift higher in claims and the +21,000 rise in continuing claims tilt modestly dovish at the margin for growth, but the stable 1.2% rate blunts the case for an abrupt repricing. Expect Treasuries to remain sensitive to follow-through data rather than this single print.
- Credit: The increase in all-programs continued weeks (+76,822 to 2,248,314) and a rise in STC/Workshare (+2,326 to 23,659) suggest a slow bleed rather than a fracture. Favor higher-quality IG over lower-rated HY, with an eye on issuers concentrated in the Northeast/West Coast.
- Equities: This is not a “recession now” signal, but cyclicals lose some altitude when continuing claims rise. Tilt toward quality growth and defensives, maintain exposure to services with pricing power, and be selective in consumer discretionary where state-level labor softness could pinch demand.
- Munis and regional exposure: Persistent elevated insured unemployment rates in RI, NJ, MA, WA, CA, IL, NY argue for credit differentiation in muni portfolios and caution in regionally concentrated small caps and regional banks.
- Macro watchlist: The weekly claims uptrend, even if shallow, raises the bar for upside payroll surprises. If the 4-week average pushes above 220k–225k and continuing claims sustain above 1.85–1.90 million, expect growth-sensitive assets to lean defensive and the curve to modestly bull-steepen.
Positioning Ideas
- Duration: Maintain or add a modest duration bias on further confirmations of labor softening; avoid chasing after a single down-week headline.
- Curve: A light bull-steepener bias makes sense if continuing claims keep grinding higher.
- Equity factor mix: Favor quality, free-cash-flow compounds, with selective defensives; underweight the most cyclical small caps tied to the Northeast/West Coast.
- Hedges: Cheap volatility hedges look attractive given state-level noise and revision risk.
The headline says “claims fell.” The data says the floor is inching up—initial claims’ 4-week average rose to 219,500, continuing claims added 21,000, and all-programs participation ticked higher. Investors shouldn’t trade the press release prose; they should trade the trajectory. The next few prints will tell us if this is drift or turn, but for now, the smart money leans quality, keeps some duration, and reads past the headline.